Statistical aspects of proficiency testing in analytical laboratories. Part 1. Ranking of participants on scores is misleading

Abstract

The general format for proficiency testing is laid down in the International Harmonised Protocol, but some statistical aspects are only implicit in that document. This paper explores one such aspect. The ranking of laboratories according to the scores obtained in a round of the proficiency test is not recommended by the Harmonised Protocol but is a popular idea, particularly amongst participants. Ranking is here shown to be seriously misleading.